João Rocha da Silva silvae86

What it is

This is my OpenCode setup for local models, mainly using a customized llama.cpp configuration.

For Qwen 3.8, DeepSeek V4, and Glimmer, the models are already trained to support reasoning effort levels. Depending on the model, these may be exposed as low, medium, high, xhigh, or as low, high, and max.

For models that support reasoning but were not trained with explicit reasoning-effort levels, such as the Qwen 3.5 and 3.6 variants, I use a token budget to limit the amount of reasoning.

Although Qwen 3.8 has built-in reasoning-effort levels, I still apply a maximum reasoning-token cap for each effort level.

Although the llama.cpp CLI flags specify preserve_thinking and a default reasoning budget, these can still be overridden through the API, so this works fine for my setup.

Thunderbolt Networking Setup on Linux

This guide documents the steps to enable and configure Thunderbolt networking on your Linux system (e.g., Ubuntu, Proxmox). It covers loading the necessary kernel modules, ensuring persistent interface naming, configuring a fixed IP address, and testing throughput with iperf3. Note that Thunderbolt 3/4 hardware advertises a raw bandwidth of 40 Gbps, but practical throughput is typically lower due to half‑duplex operation, protocol overhead, and system constraints.

Hardware Used in This Guide

This guide was developed and tested with the following hardware:

CPU: AMD Ryzen 9 PRO 6950H
- 8 cores / 16 threads
3.3 GHz base frequency / 4.94 GHz boost

2022 Zephyrus G14 Fedora Silverblue Setup

GPU Naming

Most applications in Linux will report the names of the two GPUs as follows.

Dimgrey Cavefish is dGPU
Yellow Carp is iGPU

Boot Hotkeys

Thinkfan setup

Note: I configured this thinkfan setup for my old Thinkpad w520 on Ubuntu 17.10.

1. Install necessary programs

Install lm-sensors and thinkfan.

sudo apt-get install lm-sensors thinkfan

Installing and configuring Powerline-style command line tools for developers (OSX)

Intro:

For every developer, terminal is their weapon, so why don't you customize it to become a powerful, and a beautiful weapon?

Powerline style refers to a terminal style that helps developer to keep track of their workflow easily, allows them to have perfect visual on current directories and new changes. It is also git recognizable, and failure detector that will help your development process becomes more interact and much faster.

In this guideline, I will introduce you with 2 smart shells: Zsh and Fishshell. Both are perfect for the development jobs due to its rich of resources, and user-friendly.

Note:

Installing and using schemacrawler for MacOS

A recipe for generating cool SQLite database diagrams with schemacrawler on MacOS

This was tested on MacOS 10.14.5 on 2019-07-16

schemacrawler is a free and open-source database schema discovery and comprehension tool. It can be invoked from the command-line to produce, using GraphViz, images/pdfs from a SQLite (or other database type) file. It can be used from the command-line to generate schema diagrams like these:

	#!/usr/bin/bash -xe

	cat <<EOF > "${HOME}/.config/systemd/user/zoom.slice"
	[Slice]
	AllowedCPUs=0-4
	MemoryHigh=6G
	EOF

	cat /usr/share/applications/Zoom.desktop \| sed -E 's#^(Exec=).*$#Exec=/usr/bin/systemd-run --user --slice=zoom.slice /opt/zoom/ZoomLauncher#' > "${HOME}/.local/share/applications/Zoom.desktop"

	/**
	* Download and extract `.tar.gz` file using node.js
	*
	* Created on 1400/2/7 (2021/4/27).
	* @author {@link https://mirismaili.github.io S. Mahdi Mir-Ismaili}
	*/

	import http from 'http'
	import https from 'https'
	import tar from 'tar'

	// This implementation uses any for options as the type definitions are not official so this will allow us to use hooks
	import pino from "pino";

	const hooks = {
	logMethod(inputArgs: any, method: any): any {
	if (inputArgs.length >= 2) {
	const arg1 = inputArgs.shift();
	const arg2 = inputArgs.shift();
	return method.apply(this, [arg2, arg1, ...inputArgs]);
	}

	# Setup HomeBrew: https://brew.sh/
	brew install openblas
	pip download --no-binary :all: --no-deps numpy
	unzip numpy-1.18.2.zip # (you may have newer version)
	cd numpy-1.18.2

	cat > site.cfg <<EOF
	[openblas]
	libraries = openblas
	library_dirs = $(brew --prefix openblas)/lib